girder/viime

VIIME

Getting started (Running natively)

This project includes a Pipfile to help set up a virtual environment for VIIME. To set up the virtual environment

pipenv install

and to enter it

pipenv shell

VIIME is configured from a .env file present in the current directory where it is executed. See the included .env_example for an example (or try .env_pwd, which saves DB and uploaded files in the checked out directory rather than your home directory). Once the environment is in place, you will need to initialize the tables by running

mkdir viime_sqlite
viime-cli create-tables

This will create a data directory according to the SQLALCHEMY_DATABASE_URI and UPLOAD_FOLDER environment variables defined in the .env file.

Start the development server by running:

flask run

To start the frontend, run:

cd web/
yarn
yarn serve

R Processing Functions using OpenCPU

The devops directory contains everything needed to spin up an OpenCPU instance with all dependencies necessary for the processing backend. To run it locally, build the docker container

cd devops
docker build -t viime .

and start the instance

docker run -it --rm -p 8004:8004 viime

You may find that changes to the position or size of the window in which the Docker image runs will cause the service to terminate with SIGWINCH; this is actually intended behavior and can be avoided by running the image without allocating a pseudo-terminal and keeping stdin open:

docker run --rm --name viime -p 8004:8004 viime

To stop this container, use a command like docker stop viime from another terminal.

This installation includes a custom R package called viime, which contains all functions that are exposed to the API server for CSV file processing. Functions contained in this package should accept a path to a CSV file as an argument and return a data frame. The (work-in-progress) backend code at viime/opencpu.py handles the communication between pandas and R data frames. To add additional methods exposed to the API server, add the function to the viime package and rebuild the docker image.

Database migration

This application uses flask-migrate to manage database migrations. To create a migration after changing models, run:

flask db migrate

To migrate to the latest database schema, run:

flask db upgrade

Getting started (docker-compose)

You can run the backend components for this project with Docker Compose:

cd devops/docker
docker-compose up

This command will spin up two Docker containers named docker_backend_1 and docker_opencpu_1.

Note, the web client will still need to be started using yarn serve.

Once the client is running, you can use it to upload the CSV files found in the sample data directory of this repository.